Re: binfmt_misc and persistent interpreters [fwd]

Richard Guenther (zxmpm11@student.uni-tuebingen.de)
Tue, 28 Jul 1998 11:00:32 +0200 (METDST)


Hi!

On Mon, 27 Jul 1998, Brian Jepson wrote:

> Richard,
>
> I was wondering if you had considered the possibility of extending
> binfmt_misc to work with a persistent interpreter. For example, instead
> of executing the java executable each time someone runs a java program,
> binfmt_misc would start a Java VM once, and then load the specified class
> within a thread in that VM. Essentially, there would be one Java VM
> running on a given server, and all Java processes would run within that
> VM.
>
> This approach should work for any VM that supports threading (Java, Perl),
> and the performance benefits would be fantastic. Eventually, some sort of
> load-balancing scheme could be devised, and VMs running on the same
> network could effectively provide clustering capabilities.
>
> Have you given any thought to how binfmt_misc could support this sort of
> thing, or has anyone else raised this issue? I might be interested in
> working on this, depending on how much time I have, but I would not want
> to duplicate the effort, if anyone is already doing this.

I have not even thought about that, because (initially) I wanted to
keep it as simple as possible. To make it work you have to run the
VMs as a daemon-like program that could be signalled from the kernel,
or little more complicated and uglier be notified by SYSV-messages.
Hmm... the more I think about it - it would be necessary to make
a new syscall (like request_binary_to_run) that gets the command
and args of the program/script to be executed. The daemon/VM would
call this after beeing signalled by the kernel.
This would be clean, I think (I like it - but I wont use it...).
So If you need it and think it would be really worth the effort,
you could do it like this.

Richard.

PS: forwarded to linux-kernel, if anybody has done such a think or
thinks in an other way of it.

> Thanks,
>
> Brian Jepson * (bjepson@ids.net) * http://users.ids.net/~bjepson
> Choosy mothers choose to chew Chew-Z
>

--
Richard Guenther <richard.guenther@student.uni-tuebingen.de>
PGP: 2E829319 - 2F 83 FC 93 E9 E4 19 E2 93 7A 32 42 45 37 23 57
WWW: http://www.anatom.uni-tuebingen.de/~richi/

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.altern.org/andrebalsa/doc/lkml-faq.html